Artificial intelligence: Immediate promises and ultimate goals

Abstract

Today's computers come with ‘software’ programs that are typically so difficult to use and to learn that ‘artificial intelligence’ and ‘human engineering’ can play important roles in making computers more accessible. Computers are rapidly becoming so large and powerful, and so cheap, that the large AI programs needed are becoming increasingly feasible. But the ultimate promise of artificial intelligence is far greater – to throw light on and perhaps solve some of the very difficult key problems of intelligence. It would be a tragic irony if today's fad-of-the-computer and ‘AI expert systems’ for ‘smarts’ left even fewer people trying to work on these fundamental problems.